feat(hooks): track model-skill-injector in workspace repo

hooks/model-skill-injector/HOOK.md

@@ -0,0 +1,39 @@
+---
+name: model-skill-injector
+description: "Injects model-specific best-practice context into bootstrap depending on the active LLM (Anthropic vs OpenAI)"
+metadata:
+  {
+    "openclaw": {
+      "emoji": "🧠",
+      "events": ["agent:bootstrap"],
+      "requires": { "config": ["workspace.dir"] }
+    }
+  }
+---
+
+# model-skill-injector
+
+Fires on `agent:bootstrap` and appends a model-family-specific best-practices
+hint file to the bootstrap context, so the agent always knows which tool-design
+rules apply for the currently active model — without bloating the base system
+prompt.
+
+## What it does
+
+1. Reads the active model string from the event config (`agents.defaults.model.primary`).
+2. Detects the model family by model name only:
+   - **Anthropic** — any model containing `claude`
+   - **OpenAI** — any model containing `gpt`
+3. Injects the corresponding hint file from the workspace `skills/llm-tool-best-practices/` folder as a virtual bootstrap file named `MODEL_HINTS.md`.
+4. If the model is unknown / not matched, injects a neutral fallback noting both families.
+
+## Files
+
+- Anthropic hints: `skills/llm-tool-best-practices/hints/anthropic.md`
+- OpenAI hints:    `skills/llm-tool-best-practices/hints/openai.md`
+- Fallback:        `skills/llm-tool-best-practices/hints/generic.md`
+
+## Notes
+
+- Runs read-only; never writes files.
+- Only injects the targeted section — does not load the full best-practices doc each turn.

hooks/model-skill-injector/handler.ts

@@ -0,0 +1,68 @@
+import * as fs from "fs";
+import * as path from "path";
+
+const handler = async (event) => {
+  if (event.type !== "agent" || event.action !== "bootstrap") return;
+
+  const cfg = event.context?.cfg;
+  const workspaceDir = event.context?.workspaceDir;
+  if (!workspaceDir) return;
+
+  // Resolve active model string
+  const modelRaw: string =
+    (typeof cfg?.agents?.defaults?.model === "string"
+      ? cfg.agents.defaults.model
+      : cfg?.agents?.defaults?.model?.primary) ?? "";
+
+  const model = modelRaw.toLowerCase();
+
+  // Detect model family by model name (provider-agnostic)
+  let family: "anthropic" | "openai" | "generic";
+  if (model.includes("claude")) {
+    family = "anthropic";
+  } else if (model.includes("gpt")) {
+    family = "openai";
+  } else {
+    family = "generic";
+  }
+
+  const hintsDir = path.join(
+    workspaceDir,
+    "skills",
+    "llm-tool-best-practices",
+    "hints"
+  );
+  const hintsFile = path.join(hintsDir, `${family}.md`);
+
+  let content: string;
+  try {
+    content = fs.readFileSync(hintsFile, "utf8");
+  } catch {
+    // Hints file missing — skip silently
+    console.warn(
+      `[model-skill-injector] Hints file not found: ${hintsFile} (family=${family}, model=${modelRaw})`
+    );
+    return;
+  }
+
+  // Inject as a virtual bootstrap file
+  const bootstrapFiles = event.context?.bootstrapFiles;
+  if (!Array.isArray(bootstrapFiles)) return;
+
+  // Remove any previous injection (idempotent on re-runs)
+  const existing = bootstrapFiles.findIndex((f) => f.name === "MODEL_HINTS.md");
+  if (existing !== -1) bootstrapFiles.splice(existing, 1);
+
+  bootstrapFiles.push({
+    name: "MODEL_HINTS.md",
+    content,
+    // Mark virtual so it doesn't try to read from disk
+    virtual: true,
+  });
+
+  console.log(
+    `[model-skill-injector] Injected ${family} hints (model=${modelRaw})`
+  );
+};
+
+export default handler;